In my series on the design of new voting systems, I would now like to discuss the question of high voter turnout as a goal for such systems.

Everybody agrees in enfranchisement as a goal for voting systems. Nobody eligible should find voting impossible, or even particularly hard. (And, while it may not be possible due to disabilities, it should be equally easy for a voters.)

However, there is less agreement about trading off other goals to make it trivial to vote. Some voting systems accept that there will be a certain bar of effort required to vote, and don't view it as a problem that those who will not make a certain minimum effort -- registering to vote, and coming down to a polling station -- don't vote. Other systems try to lower that bar as much as possible, with at-home voting by mail, or vote-by-internet and vote-by-phone in private elections. And many nations, such as Australia, even make voting compulsory, with fines if you don't vote.

What makes this question interesting is the numbers. With 50% voter turnouts, or even less if there is not an "interesting" race, not having trivial voting "disenfranchises" huge numbers of voters. The numbers dwarf any other number in election issues, be it more standard disenfranchisements of minorities or the disabled, or any election fraud I've ever heard about. A decision on this issue can be the most election-changing of any. Australia has 96% voter turnout, and it had 47% turnout before it passed the laws in 1924 compelling voting.

Of course, many people blanch at the idea of working towards pushing people who barely care, or actively don't care, into voting. They will, almost assuredly be minimally informed on the choices, and are surely more easily influenced into voting differently than they would if they studied the choices and acted on their own. There may be no right answer, but it is worth examining the answer built into any system. Compulsory voting remains an issue that still draws controversy even in places who have it, and there is question about whether asking people who don't care much for a choice is good or bad for the system.

There are a raft of election design factors which affect the marginal turnout:

• In many cases, the "cheapest" way to get votes for candidates is just to increase turnout by apathetic marginal supporters. It's far easier than actually changing the minds of activist voters. As such, get-out-the-vote efforts and drive-you-to-the-polls efforts are popular strategies for spending candidate money.
• Battles get fought over discouraging voting in districts which won't vote for you, to the point of active disenfranchisement.
• Voter registration plays a key role. Many programs push for things like motor-voter, or registration on election day. Some feel that loosening registration requirements is a better tradeoff than having so many people not register.
• Almost all systems allow voters who can't easily vote on election day to vote at an alternate day.
• In the extreme, many systems allow anybody to vote by mail, or require everybody to vote by mail. Oregon does the latter.
• In some locations, such as Canada, the law requires voters get 4 continuous hours off work. Some nations have a holiday on election day.
• Programs have been put in place to ease things for overseas voters, including a plan for vote-over-internet for Americans living or serving overseas. (This is more traditional enfranchisement.)
• Some election systems are slower, or break down, resulting in lines at the polls. Lines almost certainly discourage voters who see them, or hear about them through the media.

The biggest effort to increase turnout has been to abandon the traditional polling place, allowing a move to vote-at-home, most typically by mail. There is debate about the success of Oregon's project, some arguing that voter turnout is markedly up, others arguing the increase is not so great. If one accepts the former case, it becomes hard to argue for "denying" those extra voters their vote.

In particular, these efforts to increase turnout by having vote at home, or less strict registration to name a few, present compromises with other goals such as unprovable ballot, and voting security. If we like the things that voting machines can give us (better explanations for voters, error checking, access by the disabled) this suggests more radical changes, like vote-by-PC (which can be made secure, at least as far as a DRE can be secure) or vote by phone. If we're talking about tens of millions of voters these concepts can't be so readily dismissed.

Where do you think the tradeoffs lie?